Author

Chris Dunlop

A recent study published in the Proceedings of the National Academy of Sciences Nexus (Tey et al., 2024) has revealed a surprising and unnerving consequence: we are becoming tougher on the humans around us.

Researchers Kian Siong Tey, Asaf Mazar, Geoff Tomaino, Angela L. Duckworth, and Lyle H. Ungar from the University of Pennsylvania and the University of Hong Kong conducted two experiments involving over 1,200 participants, and the findings are both fascinating and concerning.

You might recognise one of the researchers Angela Duckworth from the famous book Grit - Side note, the Kindle version is only $3!

The Findings — A Negative Spillover

Essentially even after talking to an AI for a brief period of time, humans were more negative in their interactions causing a potentially “spill over effect”

The spillover effect is that after interacting with AI for a while, when you switch to talking to a human, you unintentionally carry over the harsher and more demanding tone you used with the AI. As a result, you end up being more critical and less forgiving towards the human.

Discussion

I, for one can definitely see how this would happen. Personally I do find I take a harsher tone when I am prompting, especially if the AI gets something wrong multiple times.

On Reddit there are posts of AI chat threads shutting down when people are being mean.

A week of smiley faces

After reading this study, I have spent the last week sending as many :) as I can to the AI. I have to say, it has made me feel a lot better and I think I have been a bit more positive in my interactions with people as well.

I know there is a bit of placebo going on here, but honestly, I was a grumpy as prompter before, just randomly raging at the AI a lot and I think this has helped me to be a bit more positive.

Implications of the study

I think this is one of those studies that causes you to pause and reflect on your own behavior. Personally my takeaway here is that I need to watch my tone and if I am frustrated, find a better way to channel it, than to keep raging at the AI.

My other takeaway is that for younger people growing up, this is something worth communicating to them about as well. E.G just because it’s an AI, doesn’t mean you should completely change how you interact.

This is a wild time we live in. Prior to this study I found it funny and would get a bit carried away in how I interreacted. Now I am going to be much more careful.

I never would have considered that talking in a negative way would have made me a more negative person, but hey, I’m glad I jumped onto this one earlier so I can course correct faster!